CN105096298A - Grid feature point extraction method based on fast line extraction - Google Patents
Grid feature point extraction method based on fast line extraction Download PDFInfo
- Publication number
- CN105096298A CN105096298A CN201410191199.0A CN201410191199A CN105096298A CN 105096298 A CN105096298 A CN 105096298A CN 201410191199 A CN201410191199 A CN 201410191199A CN 105096298 A CN105096298 A CN 105096298A
- Authority
- CN
- China
- Prior art keywords
- image
- point
- feature point
- line segment
- straight
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a grid feature point extraction method based on fast line extraction, and the method is used for extracting coordinates of a feature point in a grid image quickly and accurately. The execution steps are as follows: firstly carrying out the binary processing of an image, enabling an initial color image to be converted into a grey-scale map, and reducing a calculation quantity; secondly sequentially carrying out the erosion and dilation conversion and Gaussian filtering of the image, and eliminating interference and noises; carrying out the edge detection of a preprocessed image through employing a Canny edge detector, obtaining a grid contour, then carrying out fast line extraction through employing improved Hough conversion, and obtaining a line set; enabling the line set to be divided into two classes according to the slope of a line, and reducing the number of times of intersection point solving operation; enabling lines which are perpendicular to each other to be combined together so as to obtain the coordinates of intersection points, and obtaining a feature point set; setting a threshold value for the aggregation of the extracted feature points, eliminating interference, and finally obtaining a precise feature point set. The method is higher in speed and precision of feature point extraction, can eliminate the interference, and is better in fault tolerance performance.
Description
Technical field
The present invention proposes a kind of based on the Feature Points Extraction in the grid image of fast direct line drawing, be specifically related to fast direct line drawing in image and the classification of straight line, the gathering of unique point, the feature point extraction in grid fast can be realized accurately.
Background technology
The feature point extraction of image has great importance in image procossing and machine vision, has broad application prospects, also receive and pay close attention to more and more widely in fields such as Aulomatizeted Detect, biomedicine, industry manufactures.Checkerboard grid as a kind of figure of common rule, the use widely that obtains in camera calibration and mobile robot visual location.The chessboard feature point detecting method generally used now is all based on classical Harris Corner Detection Algorithm, although have stronger robustness, but this algorithm cannot adapt to the change of scale of image automatically, and the angle point utilizing this algorithm to extract is Pixel-level, time performance is good not, for chessboard, more regular, the regular image of this kind of unique point of grid, Harris algorithm is used to have obvious time waste.For grid image, maximum feature is just that its unique point intersects generation due to mutually perpendicular straight-line segment, and the unique point best approach therefore extracted in this kind of image directly utilizes straight line to ask for intersection point as unique point.In grid image, utilize straight line to carry out extract minutiae, there is obvious advantage: be first go the method finding unique point can shorten the time of feature point extraction based on fast direct line drawing, the time performance of optimized algorithm, enables this method meet the requirement of extract real-time grid search-engine point in video data; Next is the accuracy rate utilizing the method for lines detection can improve feature point extraction, and obtains the accurate coordinate of unique point; Finally, this method has very strong antijamming capability, if some unique point in grid block by environmental interference, method based on Harris Corner Detection just there will be the situation detecting unique point and omit, adopt the method for lines detection can be obtained the position of the unique point be blocked by straight-line equation simultaneous solution, draw the feature point set of complete and accurate.
In sum, based on Feature Points Extraction in the grid of fast direct line drawing, more accurate than traditional Feature Points Extraction based on Harris Corner Detection, time performance is better, interference can be eliminated simultaneously, fault freedom is better, is Feature Points Extraction in a kind of grid image of efficiently and accurately.
Summary of the invention
The invention provides Feature Points Extraction in a kind of grid of novelty, the unique point in grid can be extracted quickly and accurately, and interference can be eliminated to a certain extent, having very strong robustness, is the method for feature point extraction in a kind of grid of high efficient and reliable.
A kind of extracting method based on unique point in the grid of fast direct line drawing of the present invention, be included in the grid in true environment the straight line extracted rapidly and accurately in image, and straight line is classified, and then the unique point extracted is assembled, realize accurate feature point extraction.Particularly, the present invention includes the modules such as Image semantic classification, lines detection, straight line classification, feature point extraction, unique point gathering.Image pre-processing module: first binary conversion treatment is carried out to image, convert initial cromogram to gray-scale map, reduce the calculated amount of computing below, improve counting yield, then successively corrosion dilation transformation, gaussian filtering are carried out to image, some interference noises in removal of images; Lines detection module: utilize Canny operator to carry out rim detection to the image after pre-service, obtain the profile of grid, then utilizes the Hough transform method of improvement to carry out fast direct line drawing, obtains the set of a straight line; Straight line sort module: any two straight-line segments in grid or parallel to each other, or orthogonal, utilize this feature that obtained straight line set is divided into two classes according to slope, reduce the calculated amount of subsequent process; Feature point extraction module: mutually perpendicular line segment between two simultaneous try to achieve intersecting point coordinate, obtain the set of unique point; Unique point concentrating module: the line due to the grid in reality is not the line segment mathematically defined truly, but a band shape, this can make can extract many straight lines in a band, bring interference easily to the extraction of unique point, therefore arrange a threshold value to assemble the feature point set extracted, eliminate interference, finally obtain the set of accurate unique point.
Described is a kind of based on Feature Points Extraction in the grid of fast direct line drawing, and its workflow comprises the following steps:
Step one: binary conversion treatment is carried out to original coloured image, obtains gray-scale map;
Carrying out binary conversion treatment to image is interested part in order to retain to greatest extent in image, reduces operand below simultaneously, improves arithmetic speed.Here iteration optimal threshold algorithm determination optimal threshold is adopted: before calculating, first suppose a threshold value, then the central value of prospect under this threshold value and background is calculated, when prospect and background central value obtain mean value identical with the threshold value of supposition time, then iteration termination, and with this value for threshold value carries out binaryzation.
Step 2: 2 corrosion dilation transformations are carried out to the bianry image obtained;
Carrying out corroding conversion is frontier point in order to eliminate object, reduces target, and then can eliminate the noise spot being less than structural element; Carrying out dilation transformation is to be merged in object by all background dots with object contact, increases target, and then fills up the cavity in target.Often there is noise in real image, first corrode the operation of rear expansion, in order to noise tiny in removal of images, and be used for smooth object border.In order to strengthen the effect of stress release treatment, carry out twice corrosion expansive working.
Step 3: utilize Gaussian filter to carry out filtering to image;
Adopt the method for gaussian filtering to the smoothing process of picture signal, in post-processed, picture noise is maximum problem, and error can accumulate transmission, therefore gaussian filtering is carried out to obtain the higher image of signal to noise ratio snr (Signal/Noise, signal to noise ratio (S/N ratio)) to image.
Step 4: utilize Canny operator to carry out rim detection to the image obtained above, extract the profile of grid;
The edge of image refers to the part that image local area brightness is changed significantly.Canny operator is that rim detection effect is better than other some edge detection operators by asking the maximum value of signal function to judge image edge pixels point.
Step 5: carry out line segment extraction in profile, obtains line segment aggregate S;
It is a kind of exhaustive search strategy that common Hought converts (HT) method, and its computation complexity and space complexity are all very high, affect the speed of lines detection, therefore adopt the Hough transform method of improvement to carry out straight line rapid extraction.First in profile, carry out the Hough transform of standard, and determine that in parameter space, ballot value is not less than the straight line parameter point of ballot threshold value T by local maximum search; Then the straight line parameter point obtained is detected to each, find all unique points being not more than distance threshold w to this air line distance, constitutive characteristic point set; Finally utilize least square method to carry out fitting a straight line to feature point set, can straight line be extracted.
Step 6: according to the slope of bar straight line every in straight-line segment set, line segment aggregate S is categorized into two S set 1, S2;
Calculate the slope of the every bar straight-line segment extracted respectively, then adopt quick sorting algorithm to sort all slope value, obtain an orderly slope value sequence.Owing to only having parallel or vertical two kinds of relations between two straight lines any in grid image, therefore only have two slope value in theory.But due to can error be there is in true environment, the slope value of straight line parallel to each other can not be definitely equal, therefore arranges a threshold value, and the slope value of error in threshold range is thought equal, and straight-line segment equal for slope is put in a set, finally like this can obtain two set.
Step 7: take out straight-line segment respectively from S set 1 and S2, carries out cap between two and obtains unique point, finally obtains a feature point set;
Here, if directly ask the intersection point of two line segments as coordinate, then can produce the situation that feature point extraction is omitted, because the straight-line segment extracted is likely shorter, there is no direct intersection point with other line segment.Therefore, take simultaneous straight-line equation to solve the method for intersection point: according to slope and the starting point coordinate of every bar straight line, try to achieve the straight-line equation at each line segment place of extracting, the straight-line equation of simultaneous two straight lines solves, and can obtain the coordinate of intersection point.
Step 8: be polymerized the unique point in feature point set, exclusive PCR, obtains accurate feature point set.
Because the mesh lines in real world is banded, therefore on a mesh lines, likely can extract many straight-line segments parallel to each other, the unique point calculated also can increase to produce thereupon to be disturbed, therefore a threshold value is set, the unique point polymerization of distance in threshold range becomes a unique point, obtains accurate feature point set.
beneficial effect
The inventive method is based on fast direct line drawing, and successively carry out straight line classification, feature point extraction, unique point gathering, the final unique point extracted in grid, utilize the method accuracy rate of Corner Detection extract minutiae higher than traditional, robustness is better.And the serious interference of a part can be got rid of, even if certain unique point is sheltered from by other objects, still can try to achieve unique point by simultaneous straight-line equation, make result more accurate, fault freedom is better.
Accompanying drawing explanation
Fig. 1 method flow diagram
Fig. 2 straight line principle of classification figure
Fig. 3 unique point building up principle figure
Fig. 4 feature point extraction result figure
The unique point coordinate that Fig. 5 extracts
The anti-interference result figure of Fig. 6.
Embodiment
Fig. 1 shows process flow diagram of the present invention, is described in detail the present invention below in conjunction with drawings and Examples:
A kind of extracting method based on the unique point in the grid image of fast direct line drawing, mainly be divided into the stage that four large: (1) Image semantic classification, binary conversion treatment, corrosion dilation transformation, gaussian filtering are carried out successively to image, reduces calculated amount, some interference noises in removal of images; (2) lines detection, utilizes Canny operator to carry out rim detection to the image after pre-service, obtains the profile of grid, then utilizes least square method to carry out lines detection, obtains the set of a straight line; (3) straight line classification, is divided into two set parallel to each other the set of the straight line obtained according to the slope of straight line, reduces the calculated amount of intersection point calculation, the principle of the straight line showed in fig 2 classification; (4) feature point extraction: mutually perpendicular line segment between two simultaneous try to achieve intersecting point coordinate, obtain the set of unique point; Unique point is assembled, and arranges a threshold value and assembles the feature point set extracted, and the spacing of unique point is gathered into a unique point within threshold range, eliminates interference, figure 3 shows the principle that unique point is assembled.
Concrete implementation step is as follows:
Step 1, utilize iteration optimal threshold algorithm to carry out binary conversion treatment to input imagery, obtain gray-scale map;
Adopt the optimal threshold of iteration optimal threshold algorithm determination binaryzation operation, first suppose a threshold value before calculating, then count
The central value of the prospect under this threshold value of calculation and background, when prospect and background central value obtain mean value identical with the threshold value of supposition time, then iteration termination, and with this value for threshold value carries out binaryzation operation to image, obtain gray-scale map.
Step 2, carry out twice corrosion expansive working;
Step 3, Gaussian filter is utilized to carry out filtering;
First three step is the pretreatment stage of image, is the basis of whole feature point detecting method, is grasped by image binaryzation
Do to reduce operand; By noise tiny in corrosion expansive working removal of images, and be used for smooth object border; By gaussian filtering to image stress release treatment, prevent deviation accumulation transmission.
step 4,rim detection is carried out on the pretreated basis of previous image, obtains the profile of grid image;
Utilize Canny operator to carry out rim detection, Canny operator core is a kind of gradient method, and it uses the zero crossing of second derivative accurately to locate edge, and effect is better than other some edge detection operators.Obtain the profile accurately of grid image.
step 5,carry out fast direct line drawing;
The feature point extraction of the inventive method is the basis based on lines detection, and the speed of lines detection and the extraction of effect to unique point have a great impact.First utilize Canny operator to carry out rim detection, obtain the profile of grid, then utilize the Hough transform method of improvement to carry out fast direct line segments extraction in profile, improve the speed of lines detection, the set of the straight-line segment finally obtained is S={L1, L2 ..., Ln}.
step 6,the set of the straight line extracted is classified;
Arbitrary two straight-line segments in actual grid or parallel to each other, orthogonal, utilize this feature that the set of the straight-line segment obtained is divided into two classes according to the slope of straight line, mutually perpendicular straight-line segment is only had just to carry out find intersection computing, avoid invalid calculating, improve the efficiency of feature point extraction.Obtain two straight-line segment subclass S1={L1, L2 ..., Lm}, S2={Lm, Lm+1 ..., Ln}, arbitrary straight-line segment in S set 1 and all straight-line segments in S set 2 orthogonal.
The starting point coordinate of every bar straight-line segment and terminal point coordinate is utilized to calculate the slope of place straight line, if starting point coordinate is
, terminal point coordinate is
, then the slope k of straight line is:
(formula 1)
Here it is significant to note that, when
,
situation, now compose a larger value to replace infinity to slope k, simplify the computing method of slope.Step is as follows:
(1) according to formula (1), slope is obtained to the straight-line segment of each in set A;
(2) utilize the slope of every bar straight-line segment as mark, utilize quick sorting algorithm to sort to straight-line segment;
(3) start anew index, the turning point position of the slope value after finding sequence;
(4) straight-line segment before turning point position forms set A 1, and the straight-line segment after turning point forms set A 2, two
Straight-line segment in individual set is vertical between two arbitrarily.
step 7,utilize the set of two straight-line segments to calculate the coordinate of unique point, obtain feature point set;
If directly ask the intersection point of two line segments as coordinate, then can produce the situation that feature point extraction is omitted, because the straight-line segment extracted is likely shorter, there is no direct intersection point with other line segment.Therefore, simultaneous straight-line equation is taked to solve the method for intersection point: according to slope and the starting point coordinate of every bar straight line, try to achieve the straight-line equation at each line segment place of extracting, the straight-line equation of simultaneous two straight lines solves, the coordinate of intersection point can be obtained, avoid the situation that feature point extraction is omitted simultaneously.From S set 1, take out straight line section, and all straight-line segment simultaneous in S set 2 try to achieve intersecting point coordinate at every turn, until the straight-line segment in S1 is all taken out, obtain a feature point set V={P1, P2 ..., Pn}.The concrete method solving intersecting point coordinate is as follows:
If starting point coordinate and the terminal point coordinate of straight line 1 are respectively
,
Starting point coordinate and the terminal point coordinate of straight line 2 are respectively
,
The straight-line equation that then can obtain straight line 1 and straight line 2 is respectively:
(formula 2)
(formula 3)
Utilize formula (2) and formula (3) simultaneous solution, the intersecting point coordinate of straight line 1 and straight line 2 can be tried to achieve
for:
Step 8, to unique point and carry out unique point gathering;
Arrange a threshold value, carry out converging operation to feature point set V, the unique point of the distance between unique point in threshold range aggregates into a unique point, thus obtain accurate feature point set V '=Pk ..., Ph ..., Pm}.
Concrete unique point aggregation algorithms is as follows:
(1) each unique point is numbered, is numbered 1 ... N;
(2) arrange a stack storage feature point subset, time initial, stack is empty, arranges threshold value t;
(3) unique point being numbered 1 calculates the distance with all the other N-1 unique point successively, if the distance of certain unique point and No. 1 unique point is less than threshold value t, then numbered and changed into 1, after having calculated, all unique points being numbered 1 are formed a set, and this set is moved in stack;
(4) from remaining unique point, random selecting one is numbered the unique point of m, distance is calculated successively with all the other unique points, if the distance of certain unique point and m is less than threshold value t, then numbered and changed into m, the unique point of all m of being numbered is formed a set after having calculated, and this set is moved in stack;
(5) process is above repeated until feature point set is combined into sky;
(6) from stack, take out unique point subset successively, a random selecting unique point representatively unique point from each subset, all representative feature points form final accurate unique point set.
To sum up, based on the extracting method of the unique point in the grid image of fast direct line drawing, more accurate than the method for traditional feature point extraction based on Corner Detection, counting yield is higher, can eliminate interference, fault freedom is better, is a kind of Feature Points Extraction of efficiently and accurately simultaneously.
Claims (2)
1. based on the Feature Points Extraction in the grid image of fast direct line drawing, first pre-service is carried out to image, comprise binary conversion treatment, corrosion dilation transformation, gaussian filtering; Then carry out Canny rim detection, obtain the set of straight-line segment, and according to slope, straight-line segment is classified; Finally carry out intersection between lines point processing, obtain the set of initial characteristics point, and then the unique point extracted is assembled, obtain final unique point set, thus realize the extraction of accurate unique point.
2. according to claim 1 a kind of based on the Feature Points Extraction in the grid image of fast direct line drawing, it is characterized in that: comprise the steps:
Step 1, binary conversion treatment is carried out to original color screen table images, obtain gray-scale map;
Step 2, twice corrosion dilation transformation is carried out to the bianry image obtained;
Step 3, Gaussian filter is utilized to carry out gaussian filtering to image;
Step 4, utilize Canny operator to carry out rim detection to the image obtained above, extract the profile of grid;
Step 5, in profile, carry out line segment extraction, obtain line segment aggregate S;
In step 6, calculating line segment aggregate, the slope of every bar straight-line segment, is categorized into two S set 1, S2 line segment aggregate S;
Step 7, from S set 1 and S2, take out straight-line segment respectively, carry out cap between two and obtain unique point, finally obtain a feature point set;
Step 8, the unique point in feature point set to be polymerized, exclusive PCR, to obtain accurate feature point set.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410191199.0A CN105096298A (en) | 2014-05-08 | 2014-05-08 | Grid feature point extraction method based on fast line extraction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410191199.0A CN105096298A (en) | 2014-05-08 | 2014-05-08 | Grid feature point extraction method based on fast line extraction |
Publications (1)
Publication Number | Publication Date |
---|---|
CN105096298A true CN105096298A (en) | 2015-11-25 |
Family
ID=54576643
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410191199.0A Pending CN105096298A (en) | 2014-05-08 | 2014-05-08 | Grid feature point extraction method based on fast line extraction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105096298A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105666246A (en) * | 2016-04-12 | 2016-06-15 | 山东大学 | Tool parameter measurement device based on CCD and measurement method of device |
CN106934788A (en) * | 2015-12-30 | 2017-07-07 | 中国科学院沈阳自动化研究所 | A kind of rapid extracting method of linear angle of inclination |
CN107220939A (en) * | 2016-03-21 | 2017-09-29 | 吴健奇 | A kind of core high-spectral data geometric distortion correction method based on triangle scaling method |
CN107273809A (en) * | 2017-05-22 | 2017-10-20 | 哈尔滨工程大学 | A kind of method of the real-time autonomous classification of fishing net under water for power buoy |
CN107610506A (en) * | 2017-09-04 | 2018-01-19 | 浙江工商大学 | The detection method and system of parking position state |
CN109431511A (en) * | 2018-11-14 | 2019-03-08 | 南京航空航天大学 | A kind of human body back scoliosis angle measurement method based on Digital Image Processing |
CN109685005A (en) * | 2018-12-24 | 2019-04-26 | 中国航空工业集团公司西安飞机设计研究所 | Flight control system Data Processing Method |
CN109753766A (en) * | 2018-12-29 | 2019-05-14 | 中链科技有限公司 | Image artifacts verification information processing method and verification method based on block chain |
CN110136059A (en) * | 2019-04-04 | 2019-08-16 | 哈尔滨理工大学 | The wind electricity blade image split-joint method of linear feature |
CN110232835A (en) * | 2019-06-27 | 2019-09-13 | 浙江工业大学 | A kind of underground garage parking space detection method based on image procossing |
CN112286199A (en) * | 2020-11-06 | 2021-01-29 | 杭州国辰机器人科技有限公司 | Line patrol charging method for mobile robot |
CN112991465A (en) * | 2021-03-26 | 2021-06-18 | 禾多科技(北京)有限公司 | Camera calibration method and device, electronic equipment and computer readable medium |
CN113724313A (en) * | 2021-09-01 | 2021-11-30 | 河北工业大学 | Depth image straight line segment identification and extraction method based on correlation analysis |
-
2014
- 2014-05-08 CN CN201410191199.0A patent/CN105096298A/en active Pending
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106934788B (en) * | 2015-12-30 | 2020-11-24 | 中国科学院沈阳自动化研究所 | Rapid extraction method of straight line inclination angle |
CN106934788A (en) * | 2015-12-30 | 2017-07-07 | 中国科学院沈阳自动化研究所 | A kind of rapid extracting method of linear angle of inclination |
CN107220939A (en) * | 2016-03-21 | 2017-09-29 | 吴健奇 | A kind of core high-spectral data geometric distortion correction method based on triangle scaling method |
CN105666246B (en) * | 2016-04-12 | 2017-11-10 | 山东大学 | Cutter parameter measuring device and its measuring method based on CCD |
CN105666246A (en) * | 2016-04-12 | 2016-06-15 | 山东大学 | Tool parameter measurement device based on CCD and measurement method of device |
CN107273809A (en) * | 2017-05-22 | 2017-10-20 | 哈尔滨工程大学 | A kind of method of the real-time autonomous classification of fishing net under water for power buoy |
CN107610506A (en) * | 2017-09-04 | 2018-01-19 | 浙江工商大学 | The detection method and system of parking position state |
CN109431511A (en) * | 2018-11-14 | 2019-03-08 | 南京航空航天大学 | A kind of human body back scoliosis angle measurement method based on Digital Image Processing |
CN109431511B (en) * | 2018-11-14 | 2021-09-24 | 南京航空航天大学 | Human back scoliosis spine contour characteristic curve fitting method based on digital image processing |
CN109685005A (en) * | 2018-12-24 | 2019-04-26 | 中国航空工业集团公司西安飞机设计研究所 | Flight control system Data Processing Method |
CN109753766A (en) * | 2018-12-29 | 2019-05-14 | 中链科技有限公司 | Image artifacts verification information processing method and verification method based on block chain |
CN110136059A (en) * | 2019-04-04 | 2019-08-16 | 哈尔滨理工大学 | The wind electricity blade image split-joint method of linear feature |
CN110232835A (en) * | 2019-06-27 | 2019-09-13 | 浙江工业大学 | A kind of underground garage parking space detection method based on image procossing |
CN110232835B (en) * | 2019-06-27 | 2020-11-13 | 浙江工业大学 | Underground garage parking space detection method based on image processing |
CN112286199A (en) * | 2020-11-06 | 2021-01-29 | 杭州国辰机器人科技有限公司 | Line patrol charging method for mobile robot |
CN112991465A (en) * | 2021-03-26 | 2021-06-18 | 禾多科技(北京)有限公司 | Camera calibration method and device, electronic equipment and computer readable medium |
CN113724313A (en) * | 2021-09-01 | 2021-11-30 | 河北工业大学 | Depth image straight line segment identification and extraction method based on correlation analysis |
CN113724313B (en) * | 2021-09-01 | 2024-05-28 | 河北工业大学 | Depth image straight line segment identification and extraction method based on correlation analysis |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105096298A (en) | Grid feature point extraction method based on fast line extraction | |
CN111179324B (en) | Object six-degree-of-freedom pose estimation method based on color and depth information fusion | |
CN104217427B (en) | Lane line localization method in a kind of Traffic Surveillance Video | |
CN104463108A (en) | Monocular real-time target recognition and pose measurement method | |
CN103530590B (en) | DPM Quick Response Code recognition system | |
Wang et al. | Lane detection based on random hough transform on region of interesting | |
KR20170041168A (en) | Method, apparatus, storage medium, and device for processing lane line data | |
CN104167003A (en) | Method for fast registering remote-sensing image | |
CN105865344A (en) | Workpiece dimension measuring method and device based on machine vision | |
CN104850851A (en) | ORB feature point matching method with scale invariance | |
CN107564006B (en) | Circular target detection method utilizing Hough transformation | |
CN105574533A (en) | Image feature extraction method and device | |
CN103679193A (en) | FREAK-based high-speed high-density packaging component rapid location method | |
CN104123554A (en) | SIFT image characteristic extraction method based on MMTD | |
CN105574161A (en) | Trademark graph element identification method, device and system | |
CN104851095A (en) | Workpiece image sparse stereo matching method based on improved-type shape context | |
CN106340010A (en) | Corner detection method based on second-order contour difference | |
CN104282027A (en) | Circle detecting method based on Hough transformation | |
CN105678791A (en) | Lane line detection and tracking method based on parameter non-uniqueness property | |
CN110163894B (en) | Sub-pixel level target tracking method based on feature matching | |
CN107895166B (en) | Method for realizing target robust recognition based on feature descriptor by geometric hash method | |
Mu et al. | A novel Shi-Tomasi corner detection algorithm based on progressive probabilistic hough transform | |
CN109815763A (en) | Detection method, device and the storage medium of two dimensional code | |
CN107710229B (en) | Method, device and equipment for recognizing shape in image and computer storage medium | |
KR101391667B1 (en) | A model learning and recognition method for object category recognition robust to scale changes |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20151125 |
|
WD01 | Invention patent application deemed withdrawn after publication |